Tracking of ArUco markers using ROS, mobile robot and RealSense camera
Filip Miklaužić Object tracking with a mobile robot using computer vision
Faculty of Mechanical Engineering and Naval Architecture, Zagreb
This repository was made as an addition to the university final thesis.
Moving a mobile robot through space is a very complex task. Finding the right way and planning a trajectory can be based on different perception systems such as LIDAR, 2D and 3D vision systems. Whit a goal of recognizing and tracking a certian object, due to availability and relatively low cost, vision systems are most commonly used. The vision systems enables the perception of the working environment, while machine vision and artificial intelligence algorithms are used for its processing and interpretation. The main task of this final paper is to become familiar with the ROS system and to research, study and implement recognition and tracking algorithms on an existing mobile robot in the Computer Intelligence Laboratory. It is also necessary to design, develop and implement a planner by means of which the mobile robot will be able to maintain the reauired spatial position and orientation in relation to the object it follows.
All python scripts from this repository can be used in independent projects with minor changes Launch files that are attached are made for the specific case, but can be easily customized
ROS Melodic Morenia Ubuntu 18.04 Python 2.7
Required packages: https://github.com/pal-robotics/ddynamic_reconfigure https://github.com/IntelRealSense/realsense-ros.git https://github.com/pal-robotics/aruco_ros
-add_robot_link.py: Script that are made for creating a static transformation between camera and robot frame
-static_trans_marker: Script that are made for creating a point in space that the mobile robot must reach
-controller_positioning: Script that are made for the task of positioning in relation to the ArUco marker
-controller_following: Script that are made for following ArUco marker
-marker_positioning.launch: Script made for launch nodes for positioning in relation to the ArUco marker
-follow_marker.launch: Script made for launch nodes for following ArUco marker
-find_object_2d.launch: Script made to launch nodes for detecting objects on the scene using RGB picture
-find_object_3d.launch: Script made to launch nodes for detecting objects on the scene using RGB and PCL data
-rviz_temp_aruco_2.rviz: Template for graphical display of active topics
-Miklauzic_Filip.pdf: Final thesis